Neural Network Classification Results
4.2. Neural Network Classification Results
In the following analysis, Landsat ETM data was used as an input for neural network classification and single hidden layer architecture was applied. Kanellopoulos [25] in his study has found that the use of a single hidden layer was sufficient for most classification problems, however, once the number of inputs gets near 20, additional flexibility was required as provided by a two hidden layer network. Total system Root Mean Squared RMS error of 0.0001 was determined as a convergence point. Training was stopped when convergence was reached, or the network reached an asymptote point when training accuracy started decreasing. By default, neural network application used the equal number of hidden nodes as the number of input variable. Skidmore, et al. [28] found that the use of minimum number of hidden nodes in the neural network significantly reduced the average training accuracy, resulting in a lower accuracy of the classification result. His study found that mean training accuracy increased as more hidden nodes were added. Another study mentioned that it was sometimes useful to make the number of hidden nodes roughly equal to two or three times the total number of input classes [25]. This study used two variations of hidden nodes number, which are equal and three times of the total input number used in the neural network, while holding other parameters constant. Analysis on the classification results found that the use of more hidden nodes number in the neural network made the network architecture more complex, causing more complicated computation for training the network, which in turn needed more iterations to reach global minima. As a comparison, neural network with 7 hidden nodes reached convergence point after 5,000 iterations, whereas the use of 21 hidden nodes in the network resulted in longer training of 7,500 iterations in order to generate a similar training accuracy. Neural network was trained using back- propagation learning algorithm with learning rate and Application of Soft Classification Techniques for Forest Cover Mapping – Arief Wijaya ISSN 1858-1633 2005 ICTS 33 momentum value of 0.2 and 0.4, respectively. Learning rate reflects on the training speed, while momentum describes the sensitivity of the network to error surface. This study tried to use some variations on these parameters, and found that higher learning rate value should be balanced with the higher value of momentum, otherwise training stage became unstable and was trapped into local minima condition. According to the accuracy assessment on classification results, the best performance of neural network was achieved with 21 hidden nodes when the network was trained for 7,500 iterations.4.3. Comparison of Classification Results
Parts
» INTRODUCTION ICTS2005 The Proceeding
» Opening Fundamental Operations of Mathematical Morphology
» Morphological filter Filter theorem
» Granulometry and size distribution
» PGPC texture model and estimation of the optimal structuring element: The PGPC
» CONCLUSIONS ICTS2005 The Proceeding
» Non-ergodicity parameters RESULTS AND DISCUSSIONS 1 Partial structure factors and
» SIMULATIONS CONCLUSION ICTS2005 The Proceeding
» IMAGE RECONSTRUCTION SYSTEM DESIGN
» RESULT CONCLUSION ICTS2005 The Proceeding
» MULTI-RESOLUTION HISTOGRAM TECHNIQUE DATA
» VALIDATION STRATEGY RESULTS AND DISCUSSION
» CONCLUSION ICTS2005 The Proceeding
» INTRODUCTION DISTILATION COLUMN AND ARTIFICIAL NEURAL NETWORK
» Using Temperature Correlation Using Flow Rate Correlation
» INTRODUCTION DETECTION OF SINGLE TREE FELLING WITH SOFT
» Supervised Fuzzy c-means Method
» Neural Network classification METHOD 1. Datasets
» Neural Network Classification Results
» Comparison of Classification Results
» DISCUSSIONS ICTS2005 The Proceeding
» CONCLUSION ACKNOWLEDGEMENT ICTS2005 The Proceeding
» Caching Access List BANDWIDTH MANAGEMENT IMPLEMENTATION
» Rate Limiting BANDWIDTH MANAGEMENT IMPLEMENTATION
» BANDWIDTH MANAGEMENT CONCEPTS RESULT
» The Architecture of UML Elements Model Element
» Diagram Element Editing SYSTEM ARCHITECTURE
» Server Application Architecture Undo
» INTRODUCTION IMPLEMENTATION TESTING ICTS2005 The Proceeding
» INTRODUCTION E-PURSE ICTS2005 The Proceeding
» Interfaces Verification Tool POS – Smart Card
» MULTI AGENT SYSTEM MAS A WEIGHTED-TREE SIMILARITY ALGORITHMS
» RESULTS ICTS2005 The Proceeding
» Facial Animation Morphing and Deformation Cross Dissolve
» Feature Morphing Mesh Morphing Text-to-Speech TTS Basic Block
» Text-to-Video Algorithm Text-To-Video Stake And Desain
» Suggestion CONCLUSION AND SUGGESTION 1 Conclusion
» The Concept SHARE-IT SYSTEM ARCHITECTURE
» SHARING SCENARIO CONCLUSION ICTS2005 The Proceeding
» The Bayesian Network Model and Modified Bayesian Optimization
» Designs and Implementation SCHEDULING MODEL AND IMPLEMENTATION
» Comparison Proposed Schedule with Real Schedule
» Face-to-Face Technique Long Distance Technique
» Scenario to motivate. Context_Selection Applikasi.
» INTRODUCTION ARCHITECTURE. CONCLUSION. ICTS2005 The Proceeding
» SUGGESTION ICTS2005 The Proceeding
» Data Flow Database Structure
» EXPERIMENTAL RESULT ICTS2005 The Proceeding
» Investment Stock Prototyping System Design
» Database Model Stock Valuation
» INTRODUCTION METHODOLOGY ICTS2005 The Proceeding
» Buffer Overrun Cryptography Random Numbers
» Anti-Tampering Error Handling Injection Flaws
» Encapsulate Field Restructuring Arrays
» Generating Secure Random Number Storing Deleting Passwords
» Smart Serialization Message Digest
» Convert Message with Private Key to Public Key
» INTRODUCTION CURRENT STATUS ICTS2005 The Proceeding
» INTRODUCTION PROPOSED SIMULATION MODEL
» PARALLELIZATION STRATEGY ICTS2005 The Proceeding
» EXPERIMENTS AND DISCUSSION CONCLUSION
» INTRODUCTION RESULTS AND DISCUSSION
» EXPERIMENTAL ICTS2005 The Proceeding
» RESULT AND DISCUSSION ICTS2005 The Proceeding
» Color segmentation SYSTEM CONFIGURATION
» FEATURE CHARACTERISTICS AND GENERAL RULE
» EXPERIMENTAL RESULT CONCLUSION ICTS2005 The Proceeding
» INTRODUCTION REVIEW OF LITERATURE
» Social Economics Impact. Restructuring Impact
» Manager Application Mobile Agent Generator MAG Mobile Agents MAs
» SNMP Table Polling SNMP Table Filtering
» BREAST CARCINOMA TUMOR ICTS2005 The Proceeding
» WATERSHED ALGORITHM METHODS ICTS2005 The Proceeding
» RESULT AND DISCUSION ICTS2005 The Proceeding
» FADED INFORMATION FIELD ARCHITECTURE
» ALGORITHMS TO CHOOSE NODES TO CREATE THE FADED
» SYSTEM SIMULATIONS ICTS2005 The Proceeding
» Model and Teory MODEL, TEORY, DESIGN, IMPLEMENTATION AND
» INTRODUCTION ANALYSIS AND RESULT
» INTRODUCTION A SIMPLE MODEL OF THE QUEUING SYSTEM
» SIMULATION RESULTS DISCUSSION ICTS2005 The Proceeding
» CONCLUSION INTRODUCTION ICTS2005 The Proceeding
» Dialog Processing ADDING NONVERBAL BEHAVIOUR
» Emotion Expression Experiment ADDING NONVERBAL BEHAVIOUR
» NATURAL LANGUAGE PROCESSING EMOTION REASONING
» Fuzzy Logic Control FLC System Planning
» Digital To Analog Converter DAC Motor Driver Position Sensor Display Unit
» INTRODUCTION CONCLUSION ICTS2005 The Proceeding
» Variable-Centered Rule Structure VARIABLE-CENTERED INTELLIGENT RULE SYSTEM
» Knowledge Refinement VARIABLE-CENTERED INTELLIGENT RULE SYSTEM
» Knowledge Building VARIABLE-CENTERED INTELLIGENT RULE SYSTEM
» Knowledge Inferencing VARIABLE-CENTERED INTELLIGENT RULE SYSTEM
» INTRODUCTION BASIC CONCEPTS OF FUZZY SETS
» Calculation of the Fitness Degree
» ESTIMATING MULTIPLE NULL VALUES IN RELATIONAL
» Chen’s [6] Result This Improving Method’s Result
» The Fuzzy Set HISTOGRAM THRESHOLDING
» Fuzzy Set Similarity HISTOGRAM THRESHOLDING
» EXPERIMENTAL RESULTS ICTS2005 The Proceeding
Show more